Green’s Bronze Mystery
The Dataset (Boring and Skippable)
I will be using a dataset I already have some familiarity with. Earlier this year a #TidyTuesday data visualization event used the 120 Years of Olympic History dataset, available on Kaggle and scraped from sports-reference.com.
This dataset is missing Tokyo and Paris. I think that that’s fine. Grabbing that data from elsewhere at a matching level of detail and integrating it into this dataset is something I could have done, but chose not to do because that flavor of work is a lot less fun than what I will be doing in the rest of this report.
I will not look at Olympics from before the 2nd World War, despite them being included at a wonderful level of detail within the dataset. This little story will be focusing on visual methods, and those do not get strictly better with more data points. Those older Olympics were set in a distinctly different world and the break the Olympics took during the war offers a wonderful, unambiguous cut to split the data on.
Let’s also clear up another small matter here. My definition of Europe for the purposes of this report are the current and historic official members of the European Olympic Committees, except for the Soviet Union and Russia.
The Original Pattern
First, let’s look at these percentages among all medals mentioned above throughout our data
The golden line represents Gold medals, the silver line Silver medals and the brown-ish line Bronze medals
We will stick to visual perception here and not try to set up some proper test statistic for a difference
Bronze no longer stands out. Its line looks pretty interchangeable with Silver
Gold’s line, however, is consistently the lowest
(Though the period between 1988 and 2000 stands out for breaking that pattern a bit)
I would thus argue that Europe doesn’t have a Bronze surplus, but a Gold deficit
Theory 1: Is it a Sums thing?
Paraphrased from Hank Green:
Maybe the big number of athletes generated by summing up all the European countries helps more with Bronze than with Gold.
Gold athletes often are special generational outliers, while Bronze could be something that a group of nations can guarantee for itself just by sending enough athletes
This is actually a great intuition
Sums in statistics can have some weird properties. If you notice some weird pattern somewhere, checking whether what you are seeing one of those #JustSumThings is a great place to start your investigations
One good (and fun) way to check just that in our setting is to sum up some other arbitrary groups of countries
Here is the figure from above for countries whose official Olympics label starts with H, A, N or K
(Examples: Hungary is HUN, Argentina is ARG and Saudi Arabia is KSA)
Here we see the kind of pattern we would expect from medal percentages over the years
A turbulent mess with none of the lines particularly standing out
Here is the same for countries with neither A, B, C, D or E in their Olympic label
Same as above, just a visual mess without structure
This is the group of countries whose labels are in the correct alphabetical order
(So Denmark with DEN and Finland with FIN are included, but Sweden is not, because SWE should be ESW)
Here we see a pattern over time, but not one between the medal levels
Can you guess what country causes that recent peak?
(The answers is China with its CHN)
Next we will sum up the countries each year that only got a few medals
(Specifically 2.5% or less of the medals of each year)
This actually broadly overlaps with one of the other speculations of Hank Green in his original video
He suggests that maybe small and/or poor countries, who can only afford to support a few athletes for each game, will only invest resources into athletes which are outlier talents destined for Gold
Europe, however, can afford to support a broader base of athletes
So Europe would be represented among the outlier talents at a rate matching their population, while being over-represented among the “““average”“” Olympic-level athletes
Contrary to that theory, we see see the same pattern here as for Europe
(Note that this doesn’t fully match a grouping by wealth. All the small European countries also show up in here. Africa is plentiful, however, and even a huge country like India shows up in many years years, even in 2016.
Properly creating a subset of poor countries would have both required grabbing additional external data and been pretty awkward throughout time. Just think of the economic journey South Korea went on between fifties and our current century)
There is a neat pattern throughout time here, though. That the lines here are increasing, means that overall the most successful countries aren’t monopolizing victory as much as they did in the past
Theory 2: The Show-Offs
There are multiple avenues towards arriving at the next theory we will be looking at
For example, the short period where the pattern disappears for Europe overlaps with an interesting chunk of modern history
One might also simply wonder what is actually left after Europe and small and/or poor countries both show our suspicious pattern
And all that Gold needs to go somewhere and there are certain countries we associate with being good at Olympics
Let’s thus now look at the world’s recent great powers. Do you notice something?
China shows a surplus of Gold in recent decades
The Soviet Union shows a surplus of Gold across its entire existence with the curious exception of 64 and 68. Furthermore, look at its crazy results in 1972 and 1980 where it won around twice as much (by percentage) Gold as Bronze.
Its successor then started of with a decent Gold surplus that tapered off quickly
The US has been on a Gold Surplus for almost the entire time range, though it seems a bit as if lines grow closer after it lost its big rival
Thus I propose the following explanation for our original mystery:
A great power wanting to show off can somehow consistently catapult its talents towards Gold. And if they win Gold, other countries cannot
Thus the nations that do not belong in that category have to fight over a pool of leftover medals that is weighted towards Silver and Bronze
Now, we are already adjusting our visualizations to a medal pool with where the medal numbers are uneven
So here is a repeat of our initial Europe figure, where use the percentages to also adjust for the medals gobbled up by the great powers. These are the percentages among those leftover medals
Our pattern disappears
There is also an obvious candidate for that mysterious somehow that great powers wanting to show-off have, but other countries lack
It’s actually an interesting twist on your usual socioeconomic explanations. Being wealthy in a general sense, and letting that money trickle into sports, is neat and definitely a big advantage, but a nation that considers the Olympics a vital battlefield in its global struggle for prestige and soft power is incentivized to invest at a higher level
I propose that this probable higher investment causes the pattern that we observe
(There is probably some source out there I could use to verify the assumption that the great powers have a higher Olympics budget, but hunting that down is outside my specialty and doesn’t sound fun)
Bonus: Let’s check an alternate explanation
We could stop here, but so far I have just recycled a single figure a couple of times and that’s boring
Thus I will now do some more visualizations to check a different theory for the advantage we observe among the great powers
What else could differentiate their athletes from the rest of the world?
Many things, but what I want to look at is whether the very fact that they are collectively competing for the top spot on the medal tables has an impact on the career trajectories of the individual athletes
For the average athlete winning a Gold is probably a moment where you reevaluate your life
You have reached the peak and entered the annals of your sport
You cannot go farther than Gold
Should you really continue this challenging athlete’s life for a chance at mere tie to the victory you just experienced?
Someone from a country fighting for the overall top is part of something bigger
Obviously every athlete is competing for both themselves and their country, but that second part is a lot more ephemeral when collectively ranking 30th counts as a success
So it seems reasonable to speculate that athletes from great powers are more likely to not rest on their laurels, because they consider themselves part of a struggle that’s bigger than them
And it’s also very reasonable to assume that athletes that won once and then continued their training would be future favorites for Gold, right?
(There is also an economic layer even to this theory.
Bringing a country like Poland or Spain one of its few Gold medals will temporarily make you one of that nation’s most famous people, and thus one of its most valuable toothpaste/insurance/shoe salespeople, even if you won Shot Put or some arcane form of swimming.
The Great Powers, however, have enough winners that companies get to be picky when it comes to putting faces on their billboards.
Thus their athletes would not be busy taking advantage of their unique ooprtunity to actually monetize their career.)
Let us thus look at a figure of what happens after an athlete wins Gold
Instead of looking at each game separately, which would get too messy when looking at five categories simultaneously, I created these three rough periods. For each champion we look at their earliest Gold medal and match them into a period based on the year of that victory
Every part of my theory was apparently wrong…
(Some implementation details:
The data here excludes all team competitions because tracking an entire team is hard. Furthermore, we treat athletes participating in multiple events as multiple people. So ‘Simone Floor Biles’ would be distinct from ‘Simone Vault Biles’)
Europeans athletes more often show up for another go after their victory and champions don’t seem that much like the favorites, especially during the Cold War
I am actually not sure where I would personally put the cutoff for being a favorite, but even the slightly below 1-in-5 odds we see for modern doesn’t represent a reliable bet towards further victory
It’s tangential to my actual theory, but the difference between Modern Era and Cold War is actually pretty striking. Olympic retirement is still the top result for each group in current times, but its lead has shrunk by a lot compared to the two earlier periods
Those 80% numbers for never returning to the Olympics after a victory sure are bleak, especially considering the state-mandated drug abuse during that era
Who’s winning Gold, though, when it’s not former champions?
Let’s reverse the direction and see what the background of the athletes actually winning Gold are
(We zoom in on two specific Olympics for this, to dodge some complicated stuff in the background calculations)
Most Gold medals seem to be going to athletes participating for the first time, especially during the Cold War
2012 Europe stands out for having a pretty hefty chunk of its Gold won by people who previously participated without landing on the podium
The 2012 US and 2012 China do have former medalist as relatively major drives for Gold
We also gain see a picture that’s very different between our two snapshot, which makes me lean towards just ending this chapter with an It’s complicated
We can discard my alternate theory for 1976 due to the sky high newcomer numbers there, but it could still be plausible for 2012
Finally, let’s go for something that’s exciting and big and a bit clumsy. Let’s adjust the above figure to represent a fuller picture of these two Olympics.
(Warning: this part might look bad on some screen and browser setups)
Here we put what we already looked at on the Y-axis, and add all the other possible results on the X-axis. Note that we are looking at raw medals, rather than percentages, and thus we are back to being unable to compare Bronze counts to the other medals. I would also like to repeat another time that I needed to remove all team events for these last few figures.
With this many numbers the figure turns into a bit of a Choose-Your-Own-Adventure, though. It’s hard to pick which numbers you should be comparing.
But I believe that it’s actually a good thing to end this journey on the raw result, rather than the percentages. It’s important to remember what we are looking at here. Any athlete who made it into this data deserves to be treated as a story as well as a statistic. So here is a big, clumsy interactive figure where you can hover over points to see the person behind that result.
Go find some names you know and learn some new ones!